CSC461 Lecture 2

Download Report

Transcript CSC461 Lecture 2

CSC461 Lecture 2: Image Formation
Objectives
Fundamental imaging notions
Physical basis for image formation
 Light
 Color
 Perception
Synthetic
camera model
Other models
CSC 461: Lecture 2
1
Image Formation
In
computer graphics, we form images
which are generally two dimensional using
a process analogous to how images are
formed by physical imaging systems
 Cameras
 Microscopes
 Telescopes
 Human
CSC 461: Lecture 2
visual system
2
Elements of Image Formation
Objects
Viewer
Light
source(s)
Attributes
that govern how light interacts
with the materials in the scene
Note the independence of the objects,
viewer, and light source(s)
CSC 461: Lecture 2
3
Light
Light is the part of the electromagnetic

spectrum that causes a reaction in our
visual systems
Generally these are wavelengths in the
range of about 350-750 nm (nanometers)
Long wavelengths appear as reds and short
wavelengths as blues
CSC 461: Lecture 2
4
Ray Tracing and Geometric Optics
One
way to form an image
is to follow rays of light
from a point source
determine which rays enter
the lens of the camera.
However, each ray of light
may have multiple
interactions with objects
before being absorbed or
going to infinity.
CSC 461: Lecture 2
5
Luminance and Color Images
Luminance
 Monochromatic
 Values
are gray levels
 Analogous to working with black and white film
or television
Color
 Has
perceptional attributes of hue, saturation,
and lightness
 Do we have to match every frequency in visible
spectrum? No!
CSC 461: Lecture 2
6
Three-Color Theory
 Human
visual system
has two types of sensors
 Rods:
monochromatic,
night vision
 Cones
Color sensitive
 Three types of cone
 Only three values
(the tristimulus values) are
sent to the brain

 Need
only match these
three values
only three primary
colors
 Need
CSC 461: Lecture 2
7
Additive and Subtractive Color
Additive
color
 Form
a color by adding amounts of three
primaries
 CRTs,
projection systems, positive film
 Primaries
are Red (R), Green (G), Blue (B)
Subtractive
color
 Form
a color by filtering white light with cyan
(C), Magenta (M), and Yellow (Y) filters
 Light-material
interactions
 Printing
 Negative
CSC 461: Lecture 2
film
8
Shadow Mask CRT
CSC 461: Lecture 2
9
Pinhole Camera




xp= -x/z/d yp= -y/z/d
zp= -d

These are equations of simple
perspective
CSC 461: Lecture 2

A pinhole camera is a box
with a small hole in the
center of one side of the box
The film is placed inside the
box on one side opposite the
pinhole
The pinhole permits only a
single ray of light
Assume camera orients along
z-axis and the pinhole is at
the origin
Projection – (xp, yp, -d) is
the projection of (x, y, z)
Use trigonometry to find
projection of a point
10
Pinhole Camera




Field or angle of view – the angle made by
the largest object that the camera can
image on its film plane
Depth of field – the distance that can be
seen: from the object to the pinhole
Ideal pinhole camera – infinite depth of
field
Two disadvantages


Pinhole is too small
Camera can not be adjusted to have a different
angle of view
CSC 461: Lecture 2
11
Synthetic Camera Model
projector
•Principles:
p
image plane
projection of p
center of projection
Constructs:
•Object
•Viewer
•Light
•Film plane
CSC 461: Lecture 2
Object specification
independent of viewer
specification
Image can be computed
using simple
trigonometric calculation
The angle of view can
be changed by moving
the film plane – clipping
window
12
Advantages
Separation
of objects, viewer, light
sources
Two-dimensional graphics is a special case
of three-dimensional graphics
Leads to simple software API
 Specify
objects, lights, camera, attributes
 Let implementation determine image
Leads
CSC 461: Lecture 2
to fast hardware implementation
13
Global vs Local Lighting
Cannot
compute color
or shade of each
object independently
 Some
objects are
blocked from light
 Light can reflect from
object to object
 Some objects might
be translucent
CSC 461: Lecture 2
14
Why not ray tracing?
Ray
tracing seems more physically based
so why don’t we use it to design a graphics
system?
Possible and is actually simple for simple
objects such as polygons and quadrics with
simple point sources
In principle, can produce global lighting
effects such as shadows and multiple
reflections but is slow and not well-suited
for interactive applications
CSC 461: Lecture 2
15